13
15
# You should have received a copy of the GNU General Public License
14
16
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
19
"""A collection of function for handling URL operations."""
66
67
relpath = relpath.encode('utf-8')
67
68
# After quoting and encoding, the path should be perfectly
68
69
# safe as a plain ASCII string, str() just enforces this
69
return str(urllib.quote(relpath, safe='/~'))
70
return str(urllib.quote(relpath))
72
73
def file_relpath(base, path):
73
74
"""Compute just the relative sub-portion of a url
75
76
This assumes that both paths are already fully specified file:// URLs.
77
if len(base) < MIN_ABS_FILEURL_LENGTH:
78
raise ValueError('Length of base (%r) must equal or'
79
' exceed the platform minimum url length (which is %d)' %
80
(base, MIN_ABS_FILEURL_LENGTH))
78
assert len(base) >= MIN_ABS_FILEURL_LENGTH, ('Length of base must be equal or'
79
' exceed the platform minimum url length (which is %d)' %
80
MIN_ABS_FILEURL_LENGTH)
81
82
base = local_path_from_url(base)
82
83
path = local_path_from_url(path)
83
84
return escape(osutils.relpath(base, path))
126
scheme_end, path_start = _find_scheme_and_separator(base)
127
if scheme_end is None and path_start is None:
129
elif path_start is None:
130
path_start = len(base)
131
path = base[path_start:]
122
match = _url_scheme_re.match(base)
125
scheme = match.group('scheme')
126
path = match.group('path').split('/')
127
if path[-1:] == ['']:
128
# Strip off a trailing slash
129
# This helps both when we are at the root, and when
130
# 'base' has an extra slash at the end
133
path = base.split('/')
135
if scheme is not None and len(path) >= 1:
137
# the path should be represented as an abs path.
138
# we know this must be absolute because of the presence of a URL scheme.
140
path = [''] + path[1:]
142
# create an empty host, but dont alter the path - this might be a
143
# relative url fragment.
133
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
134
if arg_scheme_end is None and arg_path_start is None:
136
elif arg_path_start is None:
137
arg_path_start = len(arg)
138
if arg_scheme_end is not None:
140
path = arg[arg_path_start:]
141
scheme_end = arg_scheme_end
142
path_start = arg_path_start
148
match = _url_scheme_re.match(arg)
151
scheme = match.group('scheme')
152
# this skips .. normalisation, making http://host/../../..
154
path = match.group('path').split('/')
155
# set the host and path according to new absolute URL, discarding
156
# any previous values.
157
# XXX: duplicates mess from earlier in this function. This URL
158
# manipulation code needs some cleaning up.
159
if scheme is not None and len(path) >= 1:
162
# url scheme implies absolute path.
165
# no url scheme we take the path as is.
168
path = '/'.join(path)
144
169
path = joinpath(path, arg)
145
return base[:path_start] + path
170
path = path.split('/')
171
if remove_root and path[0:1] == ['']:
174
# Remove the leading slash from the path, so long as it isn't also the
175
# trailing slash, which we want to keep if present.
176
if path and path[0] == '' and len(path) > 1:
181
return '/'.join(path)
182
return scheme + '://' + '/'.join(path)
148
185
def joinpath(base, *args):
149
186
"""Join URL path segments to a URL path segment.
151
188
This is somewhat like osutils.joinpath, but intended for URLs.
153
190
XXX: this duplicates some normalisation logic, and also duplicates a lot of
181
218
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
182
219
def _posix_local_path_from_url(url):
183
220
"""Convert a url like file:///path/to/foo into /path/to/foo"""
184
file_localhost_prefix = 'file://localhost/'
185
if url.startswith(file_localhost_prefix):
186
path = url[len(file_localhost_prefix) - 1:]
187
elif not url.startswith('file:///'):
188
raise errors.InvalidURL(
189
url, 'local urls must start with file:/// or file://localhost/')
191
path = url[len('file://'):]
221
if not url.startswith('file:///'):
222
raise errors.InvalidURL(url, 'local urls must start with file:///')
192
223
# We only strip off 2 slashes
193
return unescape(path)
224
return unescape(url[len('file://'):])
196
227
def _posix_local_path_to_url(path):
239
265
This also handles transforming escaping unicode characters, etc.
241
# importing directly from ntpath allows us to test this
267
# importing directly from ntpath allows us to test this
242
268
# on non-win32 platform
243
269
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
244
270
# which actually strips trailing space characters.
245
# The worst part is that on linux ntpath.abspath has different
271
# The worst part is that under linux ntpath.abspath has different
246
272
# semantics, since 'nt' is not an available module.
250
273
win32_path = osutils._win32_abspath(path)
251
274
# check for UNC path \\HOST\path
252
275
if win32_path.startswith('//'):
267
290
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
270
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
293
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,})://(?P<path>.*)$')
271
294
_url_hex_escapes_re = re.compile(r'(%[0-9a-fA-F]{2})')
274
297
def _unescape_safe_chars(matchobj):
275
298
"""re.sub callback to convert hex-escapes to plain characters (if safe).
277
300
e.g. '%7E' will be converted to '~'.
279
302
hex_digits = matchobj.group(0)[1:]
303
326
:param url: Either a hybrid URL or a local path
304
327
:return: A normalized URL which only includes 7-bit ASCII characters.
306
scheme_end, path_start = _find_scheme_and_separator(url)
307
if scheme_end is None:
329
m = _url_scheme_re.match(url)
308
331
return local_path_to_url(url)
309
prefix = url[:path_start]
310
path = url[path_start:]
332
scheme = m.group('scheme')
333
path = m.group('path')
311
334
if not isinstance(url, unicode):
313
336
if c not in _url_safe_characters:
314
337
raise errors.InvalidURL(url, 'URLs can only contain specific'
315
338
' safe characters (not %r)' % c)
316
339
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
317
return str(prefix + ''.join(path))
340
return str(scheme + '://' + ''.join(path))
319
342
# We have a unicode (hybrid) url
320
343
path_chars = list(path)
433
456
return url_base + head, tail
436
def split_segment_parameters_raw(url):
437
"""Split the subsegment of the last segment of a URL.
439
:param url: A relative or absolute URL
440
:return: (url, subsegments)
442
(parent_url, child_dir) = split(url)
443
subsegments = child_dir.split(",")
444
if len(subsegments) == 1:
446
return (join(parent_url, subsegments[0]), subsegments[1:])
449
def split_segment_parameters(url):
450
"""Split the segment parameters of the last segment of a URL.
452
:param url: A relative or absolute URL
453
:return: (url, segment_parameters)
455
(base_url, subsegments) = split_segment_parameters_raw(url)
457
for subsegment in subsegments:
458
(key, value) = subsegment.split("=", 1)
459
parameters[key] = value
460
return (base_url, parameters)
463
def join_segment_parameters_raw(base, *subsegments):
464
"""Create a new URL by adding subsegments to an existing one.
466
This adds the specified subsegments to the last path in the specified
467
base URL. The subsegments should be bytestrings.
469
:note: You probably want to use join_segment_parameters instead.
473
for subsegment in subsegments:
474
if type(subsegment) is not str:
475
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
476
if "," in subsegment:
477
raise errors.InvalidURLJoin(", exists in subsegments",
479
return ",".join((base,) + subsegments)
482
def join_segment_parameters(url, parameters):
483
"""Create a new URL by adding segment parameters to an existing one.
485
The parameters of the last segment in the URL will be updated; if a
486
parameter with the same key already exists it will be overwritten.
488
:param url: A URL, as string
489
:param parameters: Dictionary of parameters, keys and values as bytestrings
491
(base, existing_parameters) = split_segment_parameters(url)
493
new_parameters.update(existing_parameters)
494
for key, value in parameters.iteritems():
495
if type(key) is not str:
496
raise TypeError("parameter key %r is not a bytestring" % key)
497
if type(value) is not str:
498
raise TypeError("parameter value %r for %s is not a bytestring" %
501
raise errors.InvalidURLJoin("= exists in parameter key", url,
503
new_parameters[key] = value
504
return join_segment_parameters_raw(base,
505
*["%s=%s" % item for item in sorted(new_parameters.items())])
508
459
def _win32_strip_local_trailing_slash(url):
509
460
"""Strip slashes after the drive letter"""
510
461
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
582
533
# These are characters that if escaped, should stay that way
583
534
_no_decode_chars = ';/?:@&=+$,#'
584
535
_no_decode_ords = [ord(c) for c in _no_decode_chars]
585
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
536
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
586
537
+ ['%02X' % o for o in _no_decode_ords])
587
538
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
588
539
+ [('%02X' % o, chr(o)) for o in range(256)]))
614
565
This will turn file:// urls into local paths, and try to decode
615
566
any portions of a http:// style url that it can.
617
Any sections of the URL which can't be represented in the encoding or
568
Any sections of the URL which can't be represented in the encoding or
618
569
need to stay as escapes are left alone.
620
571
:param url: A 7-bit ASCII URL
621
572
:param encoding: The final output encoding
623
:return: A unicode string which can be safely encoded into the
574
:return: A unicode string which can be safely encoded into the
624
575
specified encoding.
627
raise ValueError('you cannot specify None for the display encoding')
577
assert encoding is not None, 'you cannot specify None for the display encoding.'
628
578
if url.startswith('file://'):
630
580
path = local_path_from_url(url)
685
635
return from_location[sep+1:]
687
637
return from_location
690
def _is_absolute(url):
691
return (osutils.pathjoin('/foo', url) == url)
694
def rebase_url(url, old_base, new_base):
695
"""Convert a relative path from an old base URL to a new base URL.
697
The result will be a relative path.
698
Absolute paths and full URLs are returned unaltered.
700
scheme, separator = _find_scheme_and_separator(url)
701
if scheme is not None:
703
if _is_absolute(url):
705
old_parsed = urlparse.urlparse(old_base)
706
new_parsed = urlparse.urlparse(new_base)
707
if (old_parsed[:2]) != (new_parsed[:2]):
708
raise errors.InvalidRebaseURLs(old_base, new_base)
709
return determine_relative_path(new_parsed[2],
710
join(old_parsed[2], url))
713
def determine_relative_path(from_path, to_path):
714
"""Determine a relative path from from_path to to_path."""
715
from_segments = osutils.splitpath(from_path)
716
to_segments = osutils.splitpath(to_path)
718
for count, (from_element, to_element) in enumerate(zip(from_segments,
720
if from_element != to_element:
724
unique_from = from_segments[count:]
725
unique_to = to_segments[count:]
726
segments = (['..'] * len(unique_from) + unique_to)
727
if len(segments) == 0:
729
return osutils.pathjoin(*segments)
734
"""Extract the server address, the credentials and the path from the url.
736
user, password, host and path should be quoted if they contain reserved
739
:param url: an quoted url
741
:return: (scheme, user, password, host, port, path) tuple, all fields
744
if isinstance(url, unicode):
745
raise errors.InvalidURL('should be ascii:\n%r' % url)
746
url = url.encode('utf-8')
747
(scheme, netloc, path, params,
748
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
749
user = password = host = port = None
751
user, host = netloc.rsplit('@', 1)
753
user, password = user.split(':', 1)
754
password = urllib.unquote(password)
755
user = urllib.unquote(user)
759
if ':' in host and not (host[0] == '[' and host[-1] == ']'): #there *is* port
760
host, port = host.rsplit(':',1)
764
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
766
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
769
host = urllib.unquote(host)
770
path = urllib.unquote(path)
772
return (scheme, user, password, host, port, path)
added
removed
bzrlib/urlutils.py
